Train Wheel Bearing Failure Characteristics Of Pattern Recognition And Application Of Research

With the development of national economy, the posit ion in national economy oi railway transportation is more and more important, increasingly need to measure and diagnose with faults of key part of locomotives, so as to ensure the safe, reliable, effective and economic completion passenger and goods. And rolling bearing is one of the most important component of the railway transportation vehicle, it is one of the aptest parts damaged too, whether its operation state is normal influences directly safety of the vehicle.Rolling bearing diagnosis is limited on the spot and its signal is particular, for example special drive installation for bearing outer lane, changeful field environment etc. The early fault signal is very faint, low frequency appears more to pulse , easy to be interfered by the ambient noise even submerged. So this thesis focused on signal analysis, feature extraction with the fault diagnosing of the train rolling bearing , and seeked to more optimized methods of signal feature extraction and fault diagnosis.This thesis focused CDEFWLMS and CDRLS on signal processing and FCM on pattern recognition of fault feature in order to improve accuracy of fault diagnosis. Such adaptive algorithms based on HOS as CDEFWLMS and CDRLS can utilize the "immunity" of high-order oumulates to Gaussian noise and non-Gaussian noise, enable adaptive filter basised on HOS to work better on the spot under the strong noise background. The result is that adaptive filters based on third-order cumulants have nice qualities and higher SNR. And CDEFWLMS filter is more effective than CDRLS filter in practice, its peak value in spectrum more distinct and sharper. Through clustering these signals processed by CDEFWLMS and CDRLS by means of FCM on the spot, it is found that there is higher degree of membership for signals processed by CDEFWLMS than CDRLS. In a word, these results show both adaptive filters based third-order cumulants are helpful to bearing fault diagnosis and monitoring.This thesis is based on the development of FTRBFDI-Freight Train Rolling Bearing Fault Diagnosis Instrument. The better probation andtest, effect of FTRBFD1 on the spot further verify adaptive filtering based on HOS, ava i lahi 1 My and validity of FCM in the field of rolling bearing fault diagnosis.